LL-37 is the only cathelicidin the human body makes, and the body never builds it directly: the CAMP gene on chromosome 3p21.3 encodes an 18 kDa precursor called hCAP18, and a protease cut releases the 37-residue active peptide at the site where it is needed. The molecular account is well established in human cells and tissues. What remains genuinely contested is how much of its everyday work is bacterial killing rather than signaling, and none of this biology has been translated into an approved drug.
LL-37 is the sole human cathelicidin, encoded by the four-exon CAMP gene at chromosome 3p21.3 and liberated by proteolytic cleavage from an 18 kDa hCAP18 precursor stored in neutrophil specific granules.
The name records nothing about what the molecule does. It marks two leucine residues at the start of the mature sequence and a length of 37 amino acids, and nothing more. The part that carries the function is the charge that sequence holds, roughly +6 at physiological pH, which is what pulls the peptide onto the anionic surfaces of bacterial membranes, lipopolysaccharide and lipoteichoic acid that host cells largely lack on their outer leaflet.
LL-37 is named for the two leucine residues opening its 37-amino-acid sequence, which carries a net positive charge of roughly +6 at physiological pH from six lysines and five arginines.
CAMP is a single four-exon gene, and the split between its exons lands almost exactly on the split in the protein it makes: exons 1 through 3 carry the signal peptide and the cathelin domain, exon 4 carries the antimicrobial domain that becomes LL-37. The promoter is unusually busy for an antimicrobial peptide gene, holding a vitamin D response element alongside binding sites for C/EBP-alpha, NF-kB, AP-1 and STAT3, which is the molecular reason hormonal status, differentiation state and inflammatory signaling all move expression at once.
The isolated human cathelin-like domain does not inhibit LL-37 and unprocessed hCAP18 inhibits Gram-negative growth with efficiency comparable to the mature peptide, so the familiar account of the prodomain as a safety catch is less settled than it is usually presented.
Neutrophils are the largest source, and the common assumption that they make the peptide on demand is wrong: circulating neutrophils do not synthesize it at all. Synthesis happens in the bone marrow during granulopoiesis, at the myelocyte and metamyelocyte stages, and the mature cell leaves carrying a pre-loaded supply it will never replenish. That arrangement makes the response immediate on degranulation and also finite.
Neutrophils hold the body's largest LL-37 reservoir as hCAP18 loaded into specific granules during bone-marrow granulopoiesis, a finite supply the circulating cell never replenishes.
Activation is a single proteolytic cut, and the enzyme that makes it changes entirely with the tissue. That tissue specificity is not a footnote, because it decides which molecule actually reaches the surface: skin does not stop at LL-37, and seminal plasma produces a different peptide altogether. The lack of selectivity is worth stating plainly as well, since a free amphipathic cationic helix is cytotoxic to host cells at concentrations only modestly above its antibacterial range, which is why keeping the activating protease and its substrate in separate compartments until exocytosis reads as a safety arrangement.
| Site | Neutrophil degranulation | Skin (stratum corneum) | Seminal plasma |
|---|---|---|---|
| Enzyme | Proteinase 3, from azurophil granules | KLK5 and KLK7 | Gastricsin, also called pepsin C |
| Trigger | Exocytosis brings enzyme and substrate together outside the cell | Desquamation-linked processing | Acidic pH after ejaculation |
| Product | LL-37 | LL-37 plus shorter forms including RK-31 and KS-30 | ALL-38, LL-37 with one extra N-terminal alanine |
Proteinase 3 cleaves the alanine-leucine bond at the cathelin junction to release LL-37 when neutrophils degranulate, while skin kallikreins KLK5 and KLK7 and seminal gastricsin generate different mature forms from the same hCAP18 precursor.
The mechanism here is unusually well defined for a nutrient-to-gene link, and it is also the point where the literature most often overreaches. In human monocytes and macrophages the cell manufactures its own active hormone locally and uses it to switch on CAMP, which makes this one of the clearest vitamin D target genes in innate immunity. Cross-sectional studies linking low 25-hydroxyvitamin D to lower peptide levels are inconsistent, and supplementation trials report increases, no change, or effects that depend on baseline status, dose and tissue, so the clinical claim is not settled even though the cell biology is.
The vitamin D response element in the CAMP promoter derives from an Alu element inserted in the primate lineage and is absent from the mouse and rat genes, so rodent models cannot test this pathway at all.
Induction cannot be the front line. Transcription takes hours, while the pre-loaded neutrophil store empties in seconds to minutes, so the signals below raise the local baseline over the days after an insult rather than meeting it. They are not equally well evidenced, and the ranking carries more information than the list does.
Transcriptional induction of CAMP takes hours, so the immediate response to infection or injury comes from pre-formed neutrophil stores released on degranulation, with epithelial induction raising the local baseline over the following days.
Calling this a host defense peptide rather than an antibiotic reflects a real shift in how the field reads it, because at the concentrations present in most tissues the signaling functions are probably more consequential than the killing. Describing that signaling as anti-inflammatory is an oversimplification the data do not support: it dampens some TLR responses and amplifies others depending on cell type and stimulus.
LL-37 binds self-DNA into condensed, nuclease-resistant complexes that reach endosomal TLR9 in plasmacytoid dendritic cells, something self-DNA alone does not achieve and a mechanism central to current models of psoriasis.
The published numbers only make sense once the geography is clear, because systemic and local concentrations differ by orders of magnitude. Plasma levels sit well below what kills bacteria in a test tube, while the fluid immediately beside a degranulating neutrophil is higher than bulk sampling can measure. The peptide behaves as something that works in local hot spots, not as a circulating antibiotic.
Most commercial ELISAs raise antibodies against the precursor and report total hCAP18 rather than cleaved LL-37, so a published concentration usually describes the reservoir rather than the active peptide.
Nothing about this peptide is built to last, and the mature 37-residue form is an intermediate in a processing cascade rather than a stable endpoint. Free peptide at an inflamed site persists on the order of minutes to a few hours depending on the local protease burden, and much of what circulates is bound and inactive rather than free. The fragments are not uniformly dead: several shorter forms stay bactericidal while giving up chemotactic capacity.
Rapid proteolysis, salt and serum inhibition and a narrow window between antibacterial and cytotoxic concentrations have kept the native sequence from translating into an approved drug, pushing development toward protease-resistant analogs, D-amino acid substitutions, peptidomimetics and protected delivery.
What is conserved across every cathelicidin is the cathelin prodomain, not the antimicrobial peptide. The business end is the most divergent part of the molecule, the opposite of the usual pattern, and the family runs from amphipathic helices like LL-37 and mouse CRAMP to the proline and arginine rich extended peptides of pigs and cattle, tryptophan-rich indolicidin and disulfide-stabilized porcine protegrins. Gene number diverges too, with humans and mice carrying exactly one cathelicidin gene while cattle carry at least seven.
| Criterion | Human LL-37 | Mouse CRAMP |
|---|---|---|
| Cathelicidin genes in the genome | One (CAMP) | One (Camp) |
| Mature-region sequence identity | Reference sequence | Around 65 percent identical |
| Alu-derived vitamin D response element | Present, primate insertion | Absent |
| Vitamin D regulation of the gene | Direct transcriptional target | None reported through this element |
Mouse CRAMP shares only around 65 percent identity with LL-37 across the mature region and its gene lacks the primate Alu-derived vitamin D response element, so a mouse knockout phenotype cannot be read as a direct prediction of human biology, least of all for vitamin D.
Educational use only. This article describes what the published scientific and clinical literature reports about LL-37. It is not medical advice, and it does not recommend, prescribe, or tell anyone to use anything described here. The regulatory status shown at the top of this page reflects what the record showed on the date given there and can change. mdpep.com does not sell any substance described here, does not endorse human use of it, and does not direct anyone to obtain it.
This is not guidance for your situation. Nothing here accounts for your medical history, your current medications, or anything else specific to you, and none of it should be used to make a decision about your own health.
Affiliate disclosure. Some links on this site are affiliate links, and mdpep.com may earn a commission when they are used. That never affects what is written here, it is not an endorsement of any vendor, and it is not a statement that anything described on this page is available to buy.
Every claim here ties to a named primary source and a date, written as plain documentation with nothing for sale. That is how MD PEP covers the whole peptide market.
